修改了算法代码,并建立了一个简单的训练脚本.修改bert处理二维输入,移除PPO的permute参数

Algorithm/__init__.py

@@ -0,0 +1,27 @@
+"""
+MAGAIL Algorithm Package
+
+多智能体生成对抗模仿学习算法实现
+"""
+
+from .magail import MAGAIL
+from .ppo import PPO
+from .disc import GAILDiscrim
+from .bert import Bert
+from .policy import StateIndependentPolicy
+from .buffer import RolloutBuffer
+from .utils import Normalizer, build_mlp, reparameterize, evaluate_lop_pi
+
+__all__ = [
+    'MAGAIL',
+    'PPO',
+    'GAILDiscrim',
+    'Bert',
+    'StateIndependentPolicy',
+    'RolloutBuffer',
+    'Normalizer',
+    'build_mlp',
+    'reparameterize',
+    'evaluate_lop_pi',
+]
+